Page 1 of 1
Data Not Passing through Transformer
Posted: Fri Apr 08, 2016 7:17 am
by samyamkrishna
Hi,
The parallel job is not letting any data pass through the transformer.
There are not constraints.
There are no warnings in the director.
I am completely lost.
I read a post here regarding transformer issue between 8.5 and 9.1.
But I am designing on 8.5 and running on 8.5 too.
Regards,
Samyam
Posted: Fri Apr 08, 2016 8:06 am
by asorrell
Samyam - we are also lost. You give us no details about how the transformer is setup, so we can't help.
Posted: Fri Apr 08, 2016 8:15 am
by samyamkrishna
So the code is extracting data from DB2 into a transformer and then writing it into a seq file.
In the performance stats I can see 10 records extracted from the DB2 and 0 records after the transformer.
Posted: Fri Apr 08, 2016 10:38 am
by Thomas.B
Are you sure the lines are not written to your output file? Sometimes DataStage removes some useless stages while running.
Because the link is believed useless the performance stats are not written under it, but they are actually drag to the output.
Posted: Fri Apr 08, 2016 10:47 am
by samyamkrishna
File is empty.
Posted: Fri Apr 08, 2016 12:19 pm
by FranklinE
I believe the best help you can provide is to post the Director messages for the job. I would guess you have warnings that might help explain this.
Posted: Fri Apr 08, 2016 12:23 pm
by samyamkrishna
Nope no warning.
Posted: Fri Apr 08, 2016 2:00 pm
by chulett
I'm going to guess you have a metadata issue between the DB2 data and the flat file and they are being 'rejected' by the target stage. You might want to change the Reject Mode on it to something other than the default of 'Continue' and see if that helps shed some light.
Posted: Fri Apr 08, 2016 3:07 pm
by samyamkrishna
Hi,
yes it turned out to be a metadata issue.
specifically turned out to be NULL handling capability in the transformer.
Once we added the NULL handling function the code worked as expected.
The only surprising thing about this whole episode is there are no warning on the director.
Regards,
Samyam
Posted: Fri Apr 08, 2016 3:23 pm
by asorrell
There should have been a warning about trying to write a NULL to the sequential file. I suspect that somebody suppressed the message at the project level, not realizing that is a VERY bad precedent.
I would suggest checking out your project level message handler.
Posted: Tue Apr 19, 2016 12:51 pm
by samyamkrishna
Hi Andy,
I checked the MSG HANDLER.
There is nothing to suppress any NULL related warnings.
Regards,
Samyam
Posted: Sat Apr 23, 2016 6:12 pm
by rschirm
Find the entry in the log or the sequential output. It will state that 0 rows written and NNNN# of rows rejected.